Little racial/ethnic diversity continues to be observed among participants in clinical trials for diseases of low prevalence (<5%). The limited racial/ethnic diversity limits our ability to assess and address potential differences in therapeutic response. We propose a randomized trial of a recruitment intervention to increase racial/ethnic diversity. For more common diseases such as hypertension, community approaches have successfully increased diversity in clinical trials, but are less successful when the disease is less prevalent and persons with the disease are more difficult to contact through these targeted community approaches. Our intervention will focus on specialty clinics where treatment trials for low prevalence diseases are usually conducted. The target of the intervention will be the specialists and clinical trial coordinators. The intervention is derived from approaches to changing provider behavior in the context of improving healthcare quality (specialists) and patient navigation methods used to assist racially/ethnically diverse patients in navigating the healthcare system (coordinators). Clinical site will be the unit of randomization. Hypothesis: The proportion of racially/ethnically diverse participants enrolled in clinical sites randomized to the intervention will be greater than the proportion in sites randomized to control. We will describe mediating factors affecting success of the intervention by comparing intervention and control sites on: a) change from baseline in clinical trial specialists' and coordinators' self-efficacy and outcome expectancy in encouraging minority referrals from community physicians or their staff and in interacting with potential participants; b) performance measures (Table 2); c) changes in skills of clinical trial investigators and coordinators as measured by differences in number of referrals to the trial sites from community physicians, differences in trial participants satisfaction with clinical trial site investigators and coordinators, differences in community physicians' satisfaction with their referrals to the treatment trial, and differences in numbers and types of activities reported on recruitment logs.